[Bug 56110] mod_proxy_fcgi bogus "AH01068: Got bogus version 1" message

[bu...@apache.org]

https://issues.apache.org/bugzilla/show_bug.cgi?id=56110

--- Comment #4 from Scott Lucas <sc...@live.com> ---
(In reply to Vedran Rodic from comment #2)
> The patched version reports various values like 146, 23, 33, 244, 105 on
> similar requests, but with different file sizes (always more than 100kb
> pumped from php-fpm). 
> 
> So I guess the aborted/cancelled requests make the mod_proxy_fcgi access
> memory that shouldn't even contain the fcgi version. 
> 
> I don't know anything about the interaction between apache core, MPM event,
> mod_proxy and mod_proxy_fcgi, but it seems that it shouldn't report this
> error.
> 
> Should I open another bug? Do you need more information?

Vedran, I am still experiencing similar issues using the steps you are
mentioning to reproduce (cancelling the requests) on 2.4.10. The version is no
longer 1 of course but it isn't clear how to quash the error. Did you ever get
to the bottom of why Apache and PHP-FPM are unable to handle these cancelled
requests gracefully?

My main concern is that it represents an issue with more than just the
cancelled requests and perhaps normal requests are seeing a 503 too.

I am also getting a lot of "AH00485: scoreboard is full, not at
MaxRequestWorkers" when using PHP-FPM also. I am not sure if this issue is
related or whether there is a misconfiguration with my event MPM. I haven't
actually been able to catch the scoreboard when it was full but if it were I
don't see how it would not have reached MaxRequestWorkers and be queuing
requests.

-- 
You are receiving this mail because:
You are the assignee for the bug.

---------------------------------------------------------------------
To unsubscribe, e-mail: bugs-unsubscribe@httpd.apache.org
For additional commands, e-mail: bugs-help@httpd.apache.org